From the magazine, Microtunnelling, Tunnelling

Harker Construction rise to the challenge of Wellington contract

The company chooses to manufacture its own equipment as projects large enough to justify importing equipment are rare. Manufacturing their own gear gives Harker the opportunity to upgrade or modify their equipment as projects come up and tenders are won. They say the trick to making this work is being good at making your own gear, otherwise this can turn into a very expensive experiment.

Today Harker has six tunnel boring machines, five of them home built by sister company Tunnelling Technologies, set up in 2001 with the specific purpose of building TBMs and underground equipment. Tunnelling Technologies was established after a chance meeting between Graeme Harker and Bert Fourie, a highly qualified and experienced mechanical engineer from South Africa, internationally respected and conversant with the design and manufacture of such equipment.

The Harris Street contract, completed in November 2004 for Wellington City Council gives an insight into the company and the development of a particular machine which was designed and built specifically for the project.

The original tender called for the trenchless installation of some 630 m of 1350 mm concrete pipe at depths of 3 to 5 m, almost entirely through old reclamations completed between 1880 and 1964. The route was to extend from Wellington Harbour, through busy downtown Wellington streets to low lying areas in the town centre. Bore hole information indicated that the reclamation varied in depth between 5 and 16 m, but the majority of it was roughly 5 m thick. Beneath the reclamation was the old beach, then marine sediments, sandy silts and some gravely sands. Further into town the sediments were underlain by moderately weathered Greywacke rock.

During the tender consolidation period and after further ground investigations, Harker proposed a deeper route beneath the reclamations and old beach, through what was hoped would be more consistent and reliable ground conditions. It was known that the area had been developed and changed over the years with photographic information and historical reports indicating that old wharfs and building foundations could be in the area and old ships were reported to be buried in the reclamations. Consequently a new deep alignment was agreed with Wellington City Council that would be up to 16 m deep beneath the surface and 14 m below sea level. Over most of the line the ground was saturated and tidal in a range of 1.4 to 3 m below ground level. Hydrostatic water pressures would therefore be of concern during the machine design and it would have to be designed to withstand 2 bars of external pressure.

To make things more interesting, WCC asked Harker to take on the entire risk for natural ground conditions while Council would retain the risk for unnatural conditions. The vertical alignment was defined within a 2 m envelope within which Harker could change the levels, depending on the actual conditions encountered during the shaft constructions. Council also agreed to allow a major alignment change during the construction to eliminate some very short drives of less than 50 m if a cost benefit could be demonstrated and ground conditions permitted it. This eventually led to a new alignment called the Bond Street variation.

Also during the tender consolidation period, the machine was being considered as more ground information came available and the most desirable machine type was selected. This turned out to be a closed face slurry machine as this type of machine offers the most certainty in varied ground conditions. Furthermore, Harker already had a separation and guidance system from an earlier machine and these could be modified to suit a new machine. Harker was also aware of an old Iseki machine that had been sitting in a yard for ten years and was able to purchase it at a reasonable price. This machine was totally unsuitable as it featured old technology and was grossly underpowered, but it had a very good shell and face. These were the only salvageable parts and the rest was completely refitted. The size was 1500 mm ID so a proposal was made to WCC to increase the pipe size at no extra cost. Council accepted this and Harker had more room to work with – the new pipe size suited the old Iseki dimensions and a very good pipe that was specifically made for that machine ten years earlier by Humes Pipes.

While the machine design and manufacture was underway further ground information revealed that in the upper sections of the project from Bond to Willis Streets, ground water was negligible and it would be possible to hand jack the upper sections to get the project underway and bring in some revenue to contribute towards the machine. This also suited WCC as the disruption from operations could be kept to a minimum, with the locals already starting to complain about disruption to business. The disadvantage of this sequencing was that it meant the alignment was effectively locked in as then the rest of the job would have to join onto it.

The first drive to be undertaken was therefore a 100 m hand dug pipejack up Bond Street through marine sediments and, for the latter 30 m, Greywacke sandstone rock and progress rates of about 3 m per day were achieved.

As these works were underway the machine manufacture was being undertaken by Tunnelling Technologies. The old electric machine was gutted and fitted with new more powerful hydraulic motors, a new gear box imported from the United States, guidance system, a new face configuration and booster pumps among others. The existing separation system was completely dismantled and rebuilt in a new configuration after consulting with overseas advisers. The new system proved to be much more efficient, taking seven months to build – two months longer than anticipated – but given the seriousness of the project the machine had to be right.

Once the hand drives had been completed in Bond and Princess Streets attention was then focused on preparation for the first machine drive, a distance of 180 m between MH4 at the top of Harris Street and MH5 in Bond Street. Shaft constructions were by now starting to get very interesting. Fortunately, service relocations were tagged out of the Harker tender submission as this had become a real can of worms for Council, with many unidentified, incorrectly marked and poor condition services. These were all in the unstable, saturated tidal zone, meaning constant and significant pumping was required to keep the excavations dry.

The method of shaft construction employed was to sheetpile through the upper weaker layers into more competent ground that would seal around the bottom of the sheet piles. This worked, but not without some considerable effort. The lower sections of the shafts had to be sealed with concrete walls and cast to the ground, enveloping the lower few metres of the sheets. Excavation in these conditions was slow and the concrete sealing walls had to be cast in sections as the ground permitted. The MH4 shaft took two months to construct and was 9 m deep.

Once the shaft was completed, an entry ring was installed to launch the machine and the new machine was installed. Once the machine was launched, Harker started getting multiple faults, not with the TBM itself, but the auxiliary equipment. Individually they were minor things but the team were swamped. The faults were mainly electrical connections, power fluctuations, and hose connections – annoying things that made the first couple of weeks very frustrating for the site team. Sometimes it could take half a day to find something that took less than five minutes to fix. By the time the crew had started to get on top of these issues ground condition problems were encountered.

Where there was meant to be marine sediments an open graded stretch of cobbles was encountered. As the ground was porous, large quantities of slurry were lost into the ground and the machine’s automatic systems compensated by increasing the flows. Harker eventually overcame this problem by putting additives in the slurry, constantly topping up the slurry and installing a booster pump in the delivery line to increase the flows. Progress was slow but the machine eventually passed through these conditions.

But the fun wasn’t over as Harker started getting fragments of concrete over the shaker screens. For a short period the machine jammed and it took a day to get it moving again. Harker does not know what was encountered – nothing was meant to be there and it was assumed to be an old building foundation for which there were no records. The line passed within 5 m of the library building and once the machine had passed the building a lot of timber was encountered. This tended to choke the slurry system but, remarkably, the machine was still able to progress through it. It helped that a very high pressure water system was installed in the face that was capable of cutting timber into small pieces before passing into the grizzley (crusher). A secondary effect was that the shaker screens became bound with small timber particles and were rendered useless in less than an hours use.

Things took a turn for the worse when steel strapping started coming through the system and the machine jammed. It was released after a day but damage to the machine itself was suspected. The final 30 m of the line encountered still more timber and steel and it was suspected some timber jammed onto the side of the pipe, causing the jacking loads to spike and reach the maximum allowable for the pipes (670 tonnes). By this stage the installation had turned into a 24 hour operation, with the line being completed at the end of a particularly long shift into the early hours. Despite all the problems the average production over the line was 6 m per day with a best rate
of 24 m.

Once the machine was recovered the head was found to be almost choked with timber and steel. One large lump of steel was found in the crusher. Bearing end plays were checked and the machine was found to have sustained damage to the main bearing and gear box. Consequently the machine had to be sent back up to Auckland to be stripped down and prepared while the next drive was prepared and discussions were held with Wellington City Council to decide who should pay for the damage.

The machine took less than six weeks to fix, which was quite good considering the damage sustained. During that time the largest and deepest shaft of the project was constructed in Harris Street. By now Council had become nervous about encountering steel and started to fully appreciate the implications of having a TBM stuck 16 m beneath congested downtown streets beneath major infrastructural services. Consequently, extensive additional ground investigation work was undertaken from the Harris Street shaft using directional drilling and sonic ground penetrating radar.

Two machine drives remained to complete the project, the first being 113 m through mostly weathered rock, up Harris Street to MH 4. Eight directional drills were carried out as were three radar surveys of the line. Anomalies were found, but nothing conclusive. It was decided that since the anomalies were in rock it was unlikely that conditions like those on the previous drive would be encountered and the machine was installed.

The first section of the drive went well, but unfortunately a seal between the second and third cans blew when Harker were only half way. This led to a spectacular flooding of the line but this was brought under control by the site team and the seal was replaced underground and the seal was redesigned. The areas of anomaly were passed but nothing was encountered.

Once this drive was completed attention then focused on the last and most difficult drives to the outlet. Council carried out more directional drilling and radar testing and found four areas of significant concern. Harker proposed that the final drive be completed by hand as the ground conditions at the lower depths were found to be better than expected and sealed out the ground water. This was verified by the directional drilling which indicated very little ground water. Contingency plans were put in place which included the possibility of pressurising the face using compressed air, and the drive commenced in a 24 hour continuous operation. The drive was 140 m long and took five weeks to complete. Timber piles, steel, wire rope and an old general rubbish tip found making the decision to abandon the machine for the last drive the correct one.

During the later stages of the Harris Street contract Harker were awarded another larger project in Auckland which is currently underway and due for completion in December 2005. The Wellington machine has been skinned up to 2.5 m OD for this project and all the systems have been upgraded using previously gained knowledge, and so far the results have been very good.

Send this to a friend